Test Bench - Problem-Solving Secrets

1. Two problems?: Solve the easy one first.

Two problems at the same time can be very confusing, especially if we think it's only one problem. The indications will be scrambled. Anything we try won't seem to work. The solution is to recognize that there are two malfunctions, not one; and then to eliminate the easy one first: get it "off the table!" Now the second problem may still be difficult, but much less so.

==== A corollary: If you see something that looks even a little bit wrong, fix it and "get it off the table."

2. Be suspicious of the solitary "good guy."

One common malfunction is known as the "double select" problem. One memory location is responding all the time; not just when it is supposed to. All of the other locations are normal, but they display a muddled result. The only time we see a "good" response is when the bad cell is the one selected.

==== Might this secret apply to people? Countries??

3. Learn footprints.

Some problems seem to occur more often than others. They will have "footprints" - we will learn to recognize and fix them quickly.

4. Get help, but do some work on it first.

People like to help, and when they see our problem afresh, they will have a useful idea. Factor in the additional knowledge and background they will have. But it's a good idea to work on the problem ourself at first. We may find the answer, and we don't want to get the reputation of being a "pest".

==== In any case if we have first wrestled with the problem, the answer when it does come will "stick to the ribs."

5. Progressive redefinition of the problem.

As we home-in on the problem, we redefine it:
At first: "This machine doesn't work." Next: "Windows isn't coming up." Then: "The disk activity light isn't coming on." Finally: "Bad disk."

6. Don't ask what is wrong: Ask when it is wrong.

From #5 above, we can see the value of thinking in terms of "when, in the sequence, is it failing?" This secret is known by doctors, detectives, many others.

7. Know how the thing is supposed to work.

It has been said that all troubleshooting is based on comparison. If we compare what is supposed to happen with what is happening, we will know when the malfunction is manifesting itself, and therefore we will know the cause.

==== This secret should be good for medical research! Cure for cancer??

8. Push the problem around.

For example: If the problem seems to be "going into hiding" whenever we are looking, such as with an intermittent; maybe we can "do something" to bring it into the open, such as causing the problem to repeat itself frequently (heat, cold, software loop, etc).

9. Intermittent problem: Does anything look wrong?.

One minute the machine is working fine; the next it's sick. Back and forth. Intermittent problems are especially difficult to troubleshoot: we check the signal at one point, and it decides to measure OK, but that's a throw-off! Or, we exchange a component - seems to fix it: it's working! Then - it's broken again.

==== One solution: do a visual inspection. Look for discolorations, foreign objects, etc. Anything that looks wrong, even a little, is probably the problem. Also see #8.

10. Cut the problem in half.

There are only a finite number of components that can be bad. If we could cut that number in half enough times, we would get it down to a single one! We can cut in half time-wise: "the disk is running but we don't get the Welcome screen," or space-wise: "one of the boards is bad, or it's the chassis." With an oscilloscope: "this signal makes it to U7, but not to U8."

==== This secret doesn't work so well against intermittent problems. See #8, #9.

11. Remove candidate bad stuff one at a time.

Is the disk light bad? Check it out. It's OK? Good! Next: Is the disk running? Yes? Good! Next: (Keep doing that.) See #13.

12. Get something working.

When we're stuck and feeling discouraged, it's good psychology to know we're making some kind of progress. Grow a list of things that are working.

13. Fight!

Sometimes it seems as if we'll never find the solution. Our confidence is faltering; that makes the situation even worse. If we fight back, It has a psychological value, if nothing else.

I remember a funny parable: A frog caught in a pail of cream. He tried to jump out, but the cream was too thick. He didn't have any other ideas, so he just kept on kicking anyway. He churned the cream into butter, hopped off the top, and was on his way.

==== Also see #10, #11, & #12.

14. Take a break.

Something happens when we take a break. There may be a wrong idea stuck on our mind, a blind spot, or whatever. On returning from the break we will frequently see the solution immediately. This secret has been known to be very effective!

==== Anything from a coffee break, to lunch, to "sleep on it" should work.